|Home | About | Journals | Submit | Contact Us | Français|
Isu, the scaffold protein on which Fe-S clusters are built in the mitochondrial matrix, plays a central role in the biogenesis of Fe-S cluster proteins. We report that the reduction in the activity of several components of the cluster biogenesis system, including the specialized Hsp70 Ssq1, causes a 15–20-fold up-regulation of Isu. This up-regulation results from changes at both the transcriptional and posttranslational level: an increase in ISU mRNA levels and in stability of ISU protein. Its biological importance is demonstrated by the fact that cells lacking Ssq1 grow poorly when Isu levels are prevented from rising above those found in wild-type cells. Of the biogenesis factors tested, Nfs1, the sulfur donor, was unique. Little increase in Isu levels occurred when Nfs1 was depleted. However, its presence was required for the up-regulation caused by reduction in activity of other components. Our results are consistent with the existence of a mechanism to increase the stability of Isu, and thus its level, that is dependent on the presence of the cysteine desulfurase Nfs1.
Iron-sulfur (Fe-S) clusters are prosthetic groups required for the function of proteins in many cellular processes that involve redox reactions, enzymatic catalysis, and electron transport, as well as in some regulatory roles. Even though Fe-S clusters can be chemically reconstituted, virtually all organisms contain a highly conserved, complex assembly and transfer system that facilitates the biogenesis of Fe-S cluster–containing proteins (Barras et al., 2005 ; Johnson et al., 2005 ; Lill et al., 2006 ; Lill and Muhlenhoff, 2006 ). The central component of this system in many prokaryotes and all eukaryotes is a scaffold protein, called IscU or Isu, respectively, on which a Fe-S cluster is assembled, before transfer to a recipient apo-protein. Mitochondria of the yeast Saccharomyces cerevisiae contain two very highly conserved, functionally redundant scaffold proteins, Isu1 and Isu2, which are 83% identical and hereafter collectively referred to as Isu (Garland et al., 1999 ; Schilke et al., 1999 ). Fe-S cluster biogenesis is an essential process; either Isu1 or Isu2 must be expressed for cells to maintain viability.
Although Isu, in its role as a scaffold, is central to Fe-S cluster biogenesis, in vitro and in vivo evidence link several other proteins to the process. These include Nfs1, Yfh1, and Yah1, which are implicated in the assembly of the cluster on Isu. Nfs1, the cysteine desulfurase, acts as a sulfur donor, catalyzing the release of sulfur from cysteine, is the best established (Li et al., 1999 ; Muhlenhoff et al., 2004 ; Dutkiewicz et al., 2006 ). The iron-binding protein Yfh1, the yeast frataxin homolog, has been proposed to serve as the iron donor. In support of this idea, frataxin is capable of donating iron for Fe-S cluster reconstitution on Isu in vitro (Yoon and Cowan, 2003 ), and cells lacking Yfh1 have low Fe-S enzyme activities (Rötig et al., 1997 ). Yah1, an essential ferredoxin, also appears to play a role in cluster assembly, perhaps providing the electrons needed for the reduction of sulfur in cysteine to sulfide (Lange et al., 2000 ). Reduction in levels of Nfs1, Yfh1, or Yah1 leads to a decrease in the amount of Fe-S clusters associated with Isu, as expected for factors involved in cluster assembly (Muhlenhoff et al., 2003 ).
Proteins important in transfer of the cluster from Isu to an apo-protein include a pair of specialized molecular chaperones, the J-protein, Jac1, and its Hsp70 partner, Ssq1, which also requires the nucleotide release factor Mge1. Jac1 binds directly to Isu and targets it to Ssq1, stabilizing the Ssq1–Isu interaction by stimulating the ATPase activity of Ssq1 (Dutkiewicz et al., 2003 , 2006 ; Vickery and Cupp-Vickery, 2007 ). The chaperone pair is thought to remodel the structure of holo-Isu, facilitating transfer of the cluster to an apo-protein (Muhlenhoff et al., 2003 ; Dutkiewicz et al., 2006 ). JAC1 is an essential gene; SSQ1 is not, although cells lacking Ssq1 grow extremely poorly (Schilke et al., 1996 ). The viability of ssq1Δ cells is attributed to the ability of Ssc1, the abundant Hsp70 involved in protein translocation and folding in the mitochondrial matrix, to function with Jac1 in the Fe-S cluster biogenesis process, albeit significantly less efficiently than Ssq1 (Schilke et al., 2006 ). In addition to the chaperone system, Grx5, a nonessential monothiol glutaredoxin, plays a role in transfer of cluster from holo-Isu to an apo-protein, although its mechanism of action is not understood (Muhlenhoff et al., 2003 ). Other proteins implicated in Fe-S cluster biogenesis include the nonessential proteins Isa1, Isa2, and Nfu1, though their functions may be more specialized (Schilke et al., 1999 ; Jensen and Culotta, 2000 ; Muhlenhoff et al., 2007 ).
When yeast cells are compromised in Fe-S cluster biogenesis, not only are the activities of Fe-S proteins decreased, the so-called “iron regulon,” a set of genes regulated by the paralogous transcription factors Aft1 and Aft2, is induced (Yamaguchi-Iwai et al., 1995 ; Rutherford et al., 2001 , 2005 ; Chen et al., 2004 ; Gerber et al., 2004 ). Aft1 and Aft2 bind to a promoter sequence element designated the iron regulatory element, activating genes involved in iron uptake from the extracellular environment, as well as iron transport across organellar membranes (Garland et al., 1999 ; Rutherford et al., 2003 ). Although these two transcription factors clearly have overlapping DNA binding site specificities, they are not functionally identical. A number of genes having an iron-regulatory element in their promoters are activated by Aft1 and Aft2, but others require either Aft1 or Aft2 for activation (Rutherford et al., 2003 ; Philpott and Protchenko, 2008 ). The signal that initiates the activation of Aft1 and Aft2, and thus the iron regulon, is not known. However, several cytosolic components important in the activation of Aft1 and Aft2 have recently been identified (Ojeda et al., 2006 ; Kumanovics et al., 2008 ). One physiological outcome of the activation of the iron regulon in response to a decrease in Fe-S cluster biogenesis is an increase in mitochondrial iron levels at the expense of cytosolic iron (Lill and Muhlenhoff, 2008 ). This effect is dependent on Aft1, because mitochondrial iron levels are not increased in aft1Δ cells (Rodriguez-Manzaneque et al., 2002 ; Yang et al., 2006 ).
We previously observed that the levels of Isu are substantially higher in cells lacking the specialized Hsp70 Ssq1 than in wild-type cells (Knieszner et al., 2005 ). We have investigated the basis of this up-regulation and report that both transcriptional and posttranslational mechanisms contribute to the rise in Isu levels, with an increase in the stability of Isu protein being predominant. This increase, which is biologically important, as robust cell growth requires this higher level of Isu when chaperone activity is diminished, requires the presence of Nfs1 and is not dependent upon the increase in iron levels that occurs when cluster biogenesis is compromised.
Strains of S. cerevisiae used in this study were derived from PJ53, which is isogenic to W303: trp1-1/trp1-1 ura3-1/ura3-1 leu2-3112/leu2-3112 his3-11,15/his3-11,15 ade2-1/ade2-1 can1-100/can1-100 GAL2+/GAL2+ met2-Δ1/met2-Δ1 lys2-Δ2/lys2-Δ2. Deletions of ISU1, ISU2 (Schilke et al., 1999 ), YFH1 (Voisine et al., 2000 ), and SSQ1 (Aloria et al., 2004 ), replaced by LEU2, HIS3, HIS3, and HPH, respectively, have been described previously. Null alleles of other genes were obtained by replacing the entire coding region with the indicated marker using standard techniques: AFT2::KanMX4, GRX5::MET2, NFS1::TRP1, AFT1::LEU2, and AFT1::KanMX4. The KanMX4 cassette was obtained from the homozygous diploid yeast genomic knockout collection (Open Biosystems, Huntsville, AL; Winzeler et al., 1999 ). ISU1* were generated by site-directed mutagenesis (QuikChange protocol, Stratagene, La Jolla, CA) to change the Aft-binding site at position −245 to −239 of ISU1 from TGCACCC to TGCAGGG. A URA3-marked expression vector having the ISU1* mutation was used for a two-step disruption in isu1Δ::LEU2. The presence of ISU1* was determined by sequencing ISU1 PCR products generated from chromosomal DNA. FET3-lacZ, using published start and end points with the promoter from −506 to +3 (Li and Kaplan, 2001 ), was first constructed in pBS1-lacZ (Schilke et al., 1996 ) and then subcloned into pRS316 (Sikorski and Hieter, 1989 ) to change the marker.
ISU1 from −4 to 502 was cloned into pCM184(Gari et al., 1997 ). Strains having Isu1 under control of the tetracycline repressible promoter were created by transforming ssq1Δ aft1Δ isu1Δ isu2Δ or aft1Δ isu1Δ isu2Δ strains harboring a wild-type ISU1 URA3-marked expression vector (pRS316-Isu1; Schilke et al., 1999 ) with TRP1-marked tetR-ISU1 and counterselecting for pRS316-Isu1 on medium containing 5-flourorotic acid (5-FOA). NFS1 from 489 to 1918 was cloned into the pRS313 vector. QuikChange was used to change amino acid isoleucine 191 to serine generating nfs1I191S (Li et al., 1999 ). The construct used for generating precursor protein utilized as a control in the pulse experiments contains ISU1 from −42 to 725 in pGEM-7Zf+ (Promega, Madison, WI).
To select for suppressors of slow growth phenotype of cells lacking Ssq1, an ssq1Δ strain was mutagenized by transforming with the yeast::mTn3lacZLEU2 DNA library digested with Not I and selecting for Leu+ cells (Burns et al., 1994 ). The Leu+ transformants from each plate were resuspended in water, and 2 OD600 equivalents were spread on plates containing 5-FOA. Cells able to grow on 5-FOA were crossed to ssq1Δ carrying pRS316-SSQ1. The resulting diploids were sporulated and dissected, and the segregation of the LEU2 marker with resistance to 5-FOA was determined. Candidates that suppressed the growth defect of ssq1Δ on 5-FOA were transformed with pRSQ to rescue the DNA sequence adjacent to the transposon insertion as described (Burns et al., 1996 ). The DNA flanking the Tn3 element was sequenced and found to be an insertion after nucleotide 80 of AFT1. To verify that inactivation of AFT1 was the reason for suppression of the 5-FOA growth defect, the AFT1 open reading frame (ORF) was replaced with LEU2 by homologous recombination.
Because of slow growth and suppressor accumulation all ssq1Δ and ssq1Δ aft2Δ strains were obtained fresh from dissection of a heterozygous diploid and used immediately. Strains were crossed, sporulated and dissected to yield the various combinations described in the text. Haploid progeny were recovered and their phenotype confirmed. For analysis of the effects of Isu levels on cell growth, ssq1Δ aft1Δ isu1Δ isu2Δ tetR-ISU1 was grown overnight in rich glucose-based medium (YPD, 1% yeast extract, 2% peptone, and 2% glucose). The following day, dilutions were made into YPD with either no addition or addition of 2 μg/ml doxycycline. The doubling time was calculated by taking OD600 measurements over time, and lysates were prepared after cells had been exposed to the drug for at least 24 h. Yeast were grown on YPD or on synthetic media as described (Sherman et al., 1986 ). When indicated, the iron chelator, bathophenanthroline disulfonic acid (BPS), was added to YPD plates at 120 μM.
Gal-Nfs1 and Gal-Jac1 (Muhlenhoff et al., 2003 ) were introduced into wild type derived from PJ53 or jac1LKDDEQ strain (Andrew et al., 2006 ) by three rounds of crossing and tetrad dissection. The cells were grown overnight in galactose-based medium (1% yeast extract, 2% peptone, and 2% galactose) and then shifted to YPD for repression of GAL1 promoter.
RNA was isolated from 10- to 15-ml cultures of indicated strains by the hot phenol method (Ausubel et al., 1997 ). Ten micrograms of total RNA was separated on a 1% MOPS-formaldehyde gel, transferred to a nylon membrane, and hybridized in ULTRAhyb (Ambion, Austin, TX) overnight with radiolabeled double-stranded DNA probes. Probes were generated by random priming with [α-32P]dCTP (6000 Ci/mmol; Perkin Elmer-Cetus, Boston, MA), using the Prime-A-Gene kit (Promega), and PCR products containing the coding region of each gene. PCR products were generated using additional sequence adjacent to the gene in the plasmid. The plasmid used to generate PGK1 probe was a gift from Marvin Wickens (University of Wisconsin, Madison, WI). After stringent washes, membranes were exposed using a PhosphorImager detection screen (GE Healthcare, Piscataway, NJ), visualized on a Typhoon 941 (GE Healthcare), and the signal was quantified with ImageQuant software (Molecular Dynamics, Sunnyvale, CA). For reprobing, blots were boiled in 1% SDS and rehybridized with a new probe.
For pulse analysis four independent ssq1Δ aft1Δ isu2Δ ISU1* and 2 aft1Δ isu2Δ ISU1* transformants were treated in duplicate. 5 OD600 units of cells were grown to midlog phase, pelleted, and washed in media lacking methionine. Cells were resuspended in media lacking methionine, and 10 μl of [35S] Express Protein Labeling Mix (>1000 Ci/mmol; Perkin Elmer-Cetus) was added and incubated at 30°C for 2 min to incorporate label. These conditions were predetermined to be in the linear range of incorporated radiolabel. Cells were lysed by adding SDS lysis buffer (2% SDS, 90 mM HEPES, pH 7.5, 30 mM dithiothreitol), boiling and freezing on liquid nitrogen. Twenty microliters of ssq1Δ aft1Δ isu2Δ ISU1* and 100 μl of aft1Δ isu2Δ ISU1* were brought up to a volume of 1 ml in lysis buffer (150 mM NaCl, 50 mM HEPES, pH 7.5, 5 mM EDTA, and 1.0% Triton X-100) and centrifuged to pellet cells. The supernatant was then moved to a new tube, and an aliquot was removed and precipitated with trichloroacetic acid, for total incorporated counts. Isu1, 0.5 μl, precursor synthesized in vitro using a rabbit reticulocyte lysate system (Promega) with [35S]methionine (>800 Ci/mmol; MP Biomedicals, Solon, OH) was added to control for immunoprecipitation efficiency. Because processing of synthesized transcripts is so rapid, little or no in vivo precursor was detectable even in the case of the 2-min pulse.
Isu antibody, 1 μl, was added to the cell lysates and incubated for 2 h at 4°C with rotation. Preliminary experiments were performed with increasing concentrations of cell extracts to ensure that the antibody was sufficient for quantitative immunoprecipitations because extracts from different strains had varying concentrations of Isu. A 50% slurry of protein A beads, 20 μl, was added and incubated for 1 h at 4°C. The protein bound to the beads was washed with wash buffer (Lysis buffer + 0.1% SDS), resuspended in sample buffer, and separated by SDS-PAGE. The gel was dried and exposed to a PhophorImager detection screen, and bands were quantified with ImageQuant software. Counts were normalized for total incorporation as well as immunoprecipitation efficiency as indicated by the amount of precursor pulled down.
For pulse-chase analysis, 10 OD600 units of cells were used with 20 μl of protein labeling mix as described for pulse analysis except the radiolabeled precursor was omitted in some cases. After the pulse, cells were resuspended in 10 ml of minimal media plus 0.2 mg/ml methionine and cysteine and placed back in flasks at 30°C for the time course. At each time point cells were collected, lysed, and analyzed as described above. The zero time point was set as 100% protein, and the other time points are presented as a percentage of the zero time point. The supernatants were then reprecipitated with an Ssc1-specific antibody as a control. Data were fit to a single two-parameter exponential decay (y = ae−bx) using SigmaPlot (Systat Software, San Jose, CA).
For β-galactosidase assays, multiple independent transformants were grown in minimal media to an OD600 of 0.5–1.2, and β-galactosidase activity in Miller units was measured as described previously (Miller, 1972 ). All chemicals, unless stated otherwise, were purchased from Sigma (St. Louis, MO).
For expression of Isu2pstrep having a strep tag at the C-terminus, DNA encoding mature Isu2p (amino acid 27–156) was cloned into pET3a-Strep-tag vector (Andrew et al., 2006 ) to create pET3a-Isu2-Strep-tag. Isu2pstrep and Isu1pstrep were purified from Escherichia coli according to the protocol previously described for Isu1pstrep (Andrew et al., 2006 ).
Immunoblot analysis was carried out using the ECL detection system (GE Healthcare) according to the manufacturer's suggestion and polyclonal antibodies specific for Isa1, Isa2, Isu, and Jac1 (Voisine et al., 2001 ), Mge1 (Laloraya et al., 1994 ), and Nfs1, Nfu1, Ssc1 (Kang et al., 1990 ), Yah1 and Yfh1. The anti-Isu was raised in rabbits against mature His-tagged Isu1 protein, whereas the anti-Isa1, anti-Isa2, anti-Nfu1, and anti-Yfh1 antibodies were raised in rabbits against glutathione S-transferase (GST)-tagged protein. Anti-Nfs1 and anti-Yah1 were gifts from Andrew Dancis (University of Pennsylvania, Philadelphia, PA).
Immunoblots were conducted on 0.1 OD600 units of whole cell lysates prepared by bead beading or 10 μg of mitochondrial preparations as indicated. Mitochondria were prepared by differential centrifugation. Large quantities were prepared as previously described (Gambill et al., 1993 ). For time-course experiments, ~0.2 g of cells was harvested and lysed by bead beating in sorbitol/Tris buffer (0.6 M sorbitol, 0.1 M Tris-HCl, pH 7.5). Lysates were centrifuged at 3000 × g for 5 min to remove cell debris; the resulting supernatants were centrifuged at 16,000 × g for 10 min. The pellets were resuspended in sorbitol/Tris buffer and washed by repeated centrifugation. Final pellets were resuspended in sorbitol/Tris buffer.
As an initial step toward our goal of understanding the basis of the up-regulation of levels of Isu in the absence of Ssq1, we sought to determine the relative contributions made by Isu1 and Isu2 to the total Isu levels. First, we assessed the reactivity of antibodies raised against purified Isu1 to the 83% identical Isu2. When equivalent amounts of purified Isu1 and Isu2 were subjected to immunoblot analysis, the signals obtained were indistinguishable (Figure 1A, two left lanes). We conclude that these antibodies recognize the two proteins equally well. The ratio of the signals obtained when these antibodies were used to analyze extracts from cells expressing only Isu1 or Isu2 (isu2Δ and isu1Δ, respectively) was ~7:1. Supporting the idea that Isu1 is normally more abundant than Isu2, the signal obtained using extracts from wild-type cells was very similar to that from isu2Δ cells. Because our interest was in the up-regulation that occurs in cells lacking Ssq1, we also compared Isu levels in extracts from ssq1Δ and isu2Δ ssq1Δ. In both cases the signals were more than 15-fold higher than in extracts from cells expressing Ssq1. We conclude that Isu1 is the predominant ISU protein in wild-type cells, and its levels dramatically increase when Ssq1 is absent.
Next we asked whether ISU1 and ISU2 mRNA levels are increased in ssq1Δ cells. Such an increase would be consistent with previous observations that ISU1 and ISU2 mRNA levels increase when Aft1 is activated (Garland et al., 1999 ) and that the iron regulon is up-regulated in cells lacking Ssq1 (Knight et al., 1998 ). To determine directly whether ISU mRNA levels increase when Ssq1 is absent, we assessed ISU mRNA levels in four strains: wild type, isu1Δ, isu2Δ, and ssq1Δ. Unlike ISU1 and ISU2 proteins, the mRNAs can be distinguished, because of differences in size and divergence of nucleotide sequence (Figure 1B). ISU1 mRNA levels were very similar in wild-type and isu2Δ cells, as were the levels of ISU2 mRNA in wild-type and isu1Δ cells. However, in an SSQ1 deletion strain, both Isu1 and Isu2 mRNAs levels were increased on the order of 2–3-fold, relative to the wild-type strain. This 2–3-fold increase in ISU mRNA levels is not sufficient to explain the 15–20-fold higher level of Isu in cells lacking Ssq1 than wild-type cells.
Because the modest increase in ISU mRNA levels is insufficient to account for the high level of Isu in ssq1Δ, we wanted to further explore the basis of this regulation. However, a more comprehensive examination was hampered both by the severely compromised growth of ssq1Δ cells (Schilke et al., 1996 ) and by the propensity of the ssq1Δ strain to accumulate suppressor mutations. Therefore, we sought situations in which ssq1Δ cells grew more vigorously, whereas Isu up-regulation was maintained. We isolated insertion mutations in AFT1 as suppressors of the growth defect of ssq1Δ cells (see Materials and Methods) and subsequently determined that a complete deletion of AFT1 resulted in substantial rescue of the growth defect caused by deletion of SSQ1 (Figure 2A). As AFT1 is specifically needed to up-regulate iron transporters of the plasma and mitochondrial membranes (Yamaguchi-Iwai et al., 1995 ; Rutherford et al., 2003 ), such suppression is consistent with the idea that slow growth of ssq1Δ could partially be rescued by the depletion of iron from growth medium (Knight et al., 1998 ). To test the idea in our strain background that iron overload is in part responsible for the poor growth of Δssq1 cells, we added the iron chelator BPS to growth medium. ssq1Δ cells were able to form colonies in the presence of BPS, although growth was compromised compared with wild-type cells (Figure 2B). Also consistent with the idea that iron overload is deleterious, deletion of AFT2, which does not result in reduction of iron levels, did not suppress the growth defect of ssq1Δ cells (data not shown).
We then asked if ISU mRNA and protein was up-regulated in aft1Δ ssq1Δ cells. Similar levels of both mRNA and protein were found whether or not Aft1 was present (Figure 2, C and D). Because Aft2, the paralog of Aft1, also binds iron regulatory elements, we tested expression levels of ISU mRNA in an ssq1Δ aft2Δ strain. Isu1 and Isu2 mRNA levels were both up-regulated, consistent with the idea that either Aft1 or Aft2 is capable of activating ISU1 and ISU2. However, most relevant for this report, Isu up-regulation was maintained in aft1Δ ssq1Δ cells, making the AFT1 deletion background appropriate for our studies.
To more thoroughly assess the contribution of posttranscriptional regulation to Isu levels, we worked toward establishing conditions under which the amounts of ISU mRNA levels were the same in cells lacking or expressing Ssq1. First, we addressed the issue of transcriptional regulation of ISU1, the more highly expressed of the paralogs (Figure 3). The ISU1 promoter has a site that fits the consensus for an Aft1/2-binding site (Garland et al., 1999 ; Rutherford et al., 2003 ). We introduced a mutation changing the core consensus sequence from CACCC to CAGGG, generating the promoter mutant we designate ISU1*. Analogous mutations in other Aft1/2 target gene promoters destroy the functionality of the iron regulatory response element (Yamaguchi-Iwai et al., 1996 ). Indeed, the increase in levels of ISU1 mRNAs seen in ssq1Δ cells was obviated when the promoter mutations were present (Figure 3, compare lanes 4 and 5, top panel), indicating that Aft1/2 transcription factors are responsible for the up-regulation of Isu1 mRNA.
ISU protein levels were also compared. The levels of ISU protein were 10–15-fold higher in the absence, compared with the presence, of SSQ1 (Figure 3, lanes 1–3 compared with 4–6). Most importantly, Isu1 levels were on the order of 10-fold higher even when ISU mRNA levels were the same, that is with ISU1 under the control of the ISU1* promoter and in the absence of ISU2 (Figure 3, compare lanes 3 and 6). We conclude that substantial up-regulation of Isu1 occurs at the posttranscriptional level when Ssq1 is absent.
To begin to understand the basis of the posttranscriptional up-regulation of ISU protein, we next determined the rate of synthesis and degradation of ISU1 protein. Four independently obtained strains ssq1Δ aft1Δ isu2Δ ISU1* and two isogenic aft1Δ isu2Δ ISU1* control strains were analyzed. Cells were pulse-labeled for 2 min with [35S]methionine and [35S]cysteine. The total amount of radioactivity incorporated by the two strains was comparable, as expected, because of their similar growth rates. Extracts prepared from the strains were subjected to immunoprecipitation with Isu-specific antibodies. The rate of synthesis of Isu1 was statistically indistinguishable in cells containing or lacking Ssq1 (Figure 4A, inset).
Because there was no significant difference in the rate of synthesis of Isu1 between the two strains, we next examined the rate of Isu1 degradation. Cells were pulse-labeled and then “chased” with an excess of nonradiolabeled amino acids to prevent further incorporation of radioactivity. Extracts prepared from samples collected over a period of 4 h were subjected to immunoprecipitation, and the radioactivity was quantified. The half-life of Isu1 from the two strains differed by about fourfold, with Isu1 in cells expressing Ssq1 having a half-life of 29 min and in cells lacking Ssq1 a half-life of 111 min (Figure 4A). As a control, the half-life of the other Hsp70 of the mitochondrial matrix, Ssc1, was also determined. The half-life of Ssc1 was similar in aft1Δ isu2 Δ ISU1* and ssq1Δ aft1Δ isu2 Δ ISU1*, 112 min and 105 min, respectively (Figure 4B). We conclude that there is a significant difference in the degradation rate of Isu1 in the two strains, leading to higher levels of Isu1 in cells lacking Ssq1.
These results are consistent with the idea that the increase in stability of Isu1 is a response to an alteration in the process of Fe-S cluster biogenesis. However, it is also possible that the increase in stability of Isu is simply a consequence of its increased abundance. To test this idea, we utilized an ISU1 gene placed under the control of the tetR promoter. Using this construct, Isu1 could be expressed at the higher level typically found in ssq1Δ cells, even when Ssq1 was present (Figure 4C, inset). The rate of degradation of overexpressed Isu1 was only slightly slower in such cells compared with the rate in the cells having normal levels of Isu1, 28 min compared with 21 min (Figure 4C). We conclude that the rate of Isu1 degradation is not dependent upon its concentration in the cell.
The experiments described above establish that the levels of Isu are up-regulated when Ssq1 is absent. However, they do not answer the question of whether these higher levels are biologically important. To address this issue, we again utilized the tetR-ISU1 construct. As discussed above, Isu1 is present at levels near that found in ssq1Δ cells when expression is driven from the tetR promoter (Figure 5, right lanes). Cells grew with similar doubling times (2 h) whether Isu1 expression was driven by the endogenous promoter or by the tetR promoter.
To test the effect of lower levels of Isu1 on growth, we took advantage of the fact that addition of doxycycline represses transcription from the tetR promoter. Four independently derived transformants carrying the ssq1Δ mutation and expressing ISU1 from the tetR promoter were analyzed. In the presence of drug, expression of Isu1 was dramatically lowered in all four (Figure 5, lanes 2–5), with levels ranging from very similar to that found in the control cells containing SSQ1 to about threefold higher (lanes 3 and 2, respectively). All four grew more slowly in the presence of drug when Isu1 expression is reduced (compare lanes 2–5, and 7). However, their growth rates differed. The higher the level of Isu, the more rapid the growth rate. For example, cells having levels similar to that of the control strains expressing Isu1 from its own promoter had a doubling time of ~5 h compared with 2 h (compare lanes 1 and 3); cells having threefold higher levels grew more rapidly, with a doubling time of ~3.9 h (compare lanes 1 and 2), but still almost twice as slowly as control cells. Thus, we conclude that in the absence of Ssq1 higher levels of Isu1 protein are necessary for robust growth.
Having determined that up-regulation of Isu in the absence of the Hsp70 Ssq1 occurs at both the transcriptional and posttranslational levels, we tested the specificity of the response. First, we asked whether other components of the Fe-S cluster biogenesis pathway were also up-regulated in the absence of Ssq1. The levels of eight additional, known or putative, components of the pathway (Nfs1, Yfh1, Yah1, Mge1, Jac1, Isa1, Isa2, and Nfu1) were compared in extracts from wild-type and ssq1Δ mitochondria. None of the proteins tested differed in levels by more than twofold (Figure 6). Thus, Isu is atypical, if not unique, among components of the Fe-S cluster biogenesis pathway in regards to up-regulation in response to the absence of Ssq1.
Next we addressed the effect of reduced activity of components of the Fe-S cluster biogenesis pathway other than Ssq1 on levels of Isu. We utilized deletion mutants for testing the nonessential components Yfh1 and Grx5. To test the essential components, Jac1 and Nfs1, we used the partial loss of function mutants, nfs1L191S, a temperature-sensitive (ts) mutant (Li et al., 1999 ) and jac1LKDDEQ, which although having no obvious growth defect, encodes a protein partially defective in binding to Isu (Andrew et al., 2006 ). Substantially higher levels of Isu were observed in grx5Δ and yfh1Δ, as well as jac1LKDDEQ, but not in nfs1L191S (Figure 7A). Because the difference between the jac1LKDDEQ and nfs1L191S was striking, we considered the possibility that the NFS1 mutant had less of an effect on Fe-S cluster biogenesis than the JAC1 mutant.
The degree of up-regulation of the “iron regulon” is an indication of the severity of the defect in Fe-S cluster biogenesis. We therefore monitored the response by testing expression from the promoter of FET3, which encodes the high-affinity plasma membrane iron transporter whose expression is driven by Aft1 and Aft2, utilizing a FET3:lacZ fusion. β-Galactosidase activity was threefold higher in nfs1L191S mutant compared with the control strain, but the jac1LKDDEQ mutant and its control strain had very similar activities (Figure 7B). Thus, all available data indicate that the NFS1 mutant has greater effects on Fe-S cluster biogenesis than the JAC1 mutant.
To determine whether the increase in Isu levels in jac1LKDDEQ cells is due to an increase in stability compared with cells expressing wild-type Jac1, we carried out a pulse-chase analysis of protein degradation. The half-life of Isu1 from the two strains differed by ~5.4-fold, with Isu1 in cells expressing wild-type Jac1 having a half-life of 36 min and in cells expressing Jac1LKDDEQ, a half-life of 197 min (Figure 7C). As expected, the level of Isu1 mRNA was similar in both strains (data not shown). Thus, Isu1 is regulated at the level of protein degradation in cells having reduced Ssq1 or Jac1 activity.
To extend the comparison between Jac1 and Nfs1, we placed the expression of these proteins under the control of the GAL1 promoter, allowing us to monitor Isu levels after repression of either Jac1 or Nfs1 upon transfer from galactose- to glucose-based medium. In the case of Jac1, the levels of Isu1 began increasing when Jac1 levels reached 25% of normal, rising to ~15-fold higher 6 h later (Figure 8A). Nfs1 depletion showed a very different pattern. About 12 h after Nfs1 dropped to below 25% of normal levels, Isu levels began to rise, but only about fourfold, and dropped to about twofold by the end of the experiment (Figure 8A). In sum, a decrease in Nsf1 activity did not result in a substantial increase in Isu levels.
Of the five components tested for the effect of their activity on the expression of Isu, only Nfs1 depletion failed to show a 10-fold or greater increase in levels. Therefore, we decided to ask whether Nfs1 was required for the increase in Isu levels to occur. To this end we combined the jac1LKDDEQ mutation with the Gal:Nfs1 repression system. In the galactose-based medium, conditions under which expression of Nfs1 was not repressed, Isu levels were high, as expected. However, after inhibition of Nfs1 synthesis the level of Isu dropped (Figure 8B). Our results are consistent with Nfs1 being necessary for the increase in Isu levels.
We report that the cellular level of the Fe-S cluster scaffold Isu is regulated both at the transcriptional and the posttranslational levels, that is via changes in the rates of mRNA synthesis and protein degradation. The rise in protein stability accounts for the majority of the increase in Isu levels. Upregulation occurs in most, but not all circumstances in which mitochondrial Fe-S cluster biogenesis is compromised. This up-regulation of Isu is biologically important when Fe-S cluster biogenesis is compromised, because its prohibition compromises cell growth.
Both the transcriptional and posttranslational regulation is apparently unique to Isu among the known biogenesis factors, as no significant increase in the levels of the eight other components of the Fe-S biogenesis system tested was observed. In the case of the transcriptional response, this conclusion is supported by additional information. The promoters of both ISU1 and ISU2 contain a consensus binding site for the Aft1 and Aft2 transcription factors, whereas none of the other genes encoding the components we tested have such a site. Nor have they been shown to be up-regulated in genome-wide microarray studies when Aft1/2 are activated, for example, in the case of the deletion of YFH1 (Foury and Talibi, 2001 ) or depletion of YAH1 (Hausmann et al., 2008 ). Perhaps Isu's uniqueness is not surprising, because it is the scaffold on which the clusters are assembled and thus is the central player in the process of Fe-S cluster biogenesis.
How the increased stability of Isu we observed relates to this pathway of Aft1/2 activation remains open. However, two pieces of evidence indicate that the increased stability of Isu does not depend on increased mitochondrial iron levels, one of the outcomes of Aft activation. First, the jac1LKDDEQ mutant mounts a robust response, but iron levels are not increased in its mitochondria. Second, the increase in mitochondrial iron levels does not occur in cells lacking Aft1 (Rodriguez-Manzaneque et al., 2002 ; Yang et al., 2006 and data not shown). However, large increases in Isu levels can occur in the absence of Aft1. Conversely, the increase in mitochondrial iron levels does not depend on Nfs1, as it was previously shown that a decrease in Nfs1 activity leads to an increase in mitochondrial iron levels (Li et al., 1999 ).
What is the mechanism behind the regulation of Isu stability? The idea that the increase in protein stability is simply an indirect result of the 3- to 4-fold increase of protein concentration due to the mRNA up-regulation is ruled out: The stability of Isu was the same in a wild-type strain with Isu at normal levels or when overexpressed 15-fold. Nor do our results suggest that the increase correlates with either the efficiency of cluster formation or cluster transfer from Isu, as absence of components involved in both results in increased Isu protein levels. High levels of Isu were first observed in cells lacking Ssq1, which is important for cluster transfer (Dutkiewicz et al., 2006 ; Knieszner et al., 2005 ). However, an increase in Isu levels was also observed in the absence of Yfh1, the putative iron donor for cluster formation. Indeed, Lill and coworkers have shown that upon depletion of Yfh1, cluster formation on Isu is greatly diminished (Gerber et al., 2003 ).
Perhaps the best clue as to mechanism, is our finding that of the five components of the Fe-S biogenesis system tested, decrease in activity of four resulted in a robust increase in Isu levels. The outlier was Nfs1, whose depletion did not result in a substantial increase in Isu levels. More importantly however, Nfs1 was required for a JAC1 mutant to sustain high Isu levels. Nfs1, a cysteine desulfurase, serves as the sulfur donor for cluster synthesis, which is thought to be the first step in cluster synthesis (Smith et al., 2001 ). Whether Nfs1 binding to Isu fosters a conformation of Isu that is required for stabilization or whether its activity or presence is required to transmit a signal will require further exploration.
Whatever the exact mechanism behind the increase in stability of Isu, it is important to note that the increase is important for cells to maintain robust growth in the absence of the specialized Hsp70 Ssq1. Although we found this response because of our analysis of the effect of mutations that compromised the Fe-S biogenesis system, it likely occurs normally when yeast are exposed to environmental conditions that either limit Fe-S cluster formation or affect cluster stability, such as oxidative stress. As a regulatory mechanism, increase in protein stability has an advantage of potentially being rapid, taking less time and less expenditure of energy than an increase in transcription that requires not only mRNA transcription but translation and protein translocation as well. Interestingly cells have evolved not one, but at least two, mechanisms to up-regulate Isu, underscoring the importance of maintaining the functionality of the essential and multifunctional proteins having Fe-S clusters.
We thank Andrew Dancis for Nsf1- and Yah1-specific antibodies, Roland Lill (University of Marburg) for GAL-NFS1 and GAL-JAC1; and Jarek Marsalek and Amy Prunuske for helpful comments on the manuscript. This work was supported by the National Institutes of Health Grant GM278709 and the Muscular Dystrophy Association (E.A.C.).
This article was published online ahead of print in MBC in Press (http://www.molbiolcell.org/cgi/doi/10.1091/mbc.E08-06-0622) on October 8, 2008.